Machine-Learning for Training
Last updated
Last updated
Copyright Spielworks GmbH 2024
Chainmonsters presents a stellar and truly unique approach to incorporating state of the art technologies like Machine Learning with exhilarating gameplay.
We are putting the player and their individual Chainmons back into action by introducing smart and continuously learning Chainmons that can be adjusted, enhanced and specialized by players tailored around their own unique playstyle and preference.
Every Chainmon has a unique nature that pre-determines their initial intelligence, ability to act and adjust within current events like battling, farming and other activities.
Through the player's actions, unique playstyle, training and optimized strategies, the Chainmons enhance their capabilities over time into either highly specialized variants or all-rounders suitable for any task.
When Chainmonsters launches, each species has a base intelligence associated through their unique nature.
There are many variations of base level intelligence so depending on your luck, progress in the game and other factors you may capture a variation that may be further or lesser advanced than their overall peers or even have a solid foundation in more specialized activities already.
This aspect is fairly balanced but still enables a level of variation that further boosts each individual's uniqueness.
The Training section outlines the different opportunities to train and utilize the Chainmons in much more detail so we will focus on the underlying high-level concepts a bit more here:
Any action a player takes when interacting with a specific Chainmon influences their overall learning and improvement aspects over time.
Chainmonsters utilizes a mix of imitation and autonomous training to enable both truly unique and customizable experiences depending on each player.
For example, the player actively controls their Chainmons in battle against opponents which generates training data used for imitation of this particular Chainmon against this specific opponent. While initially, the player-controlled capabilities and optimization might be more effective than the pre-trained model - of course also depending on the player's skills and deep understanding of the game mechanics - with more training and experience the Chainmon aims to become superior and can take on hard encounters and boss battles fully autonomously.
Chainmonsters - as an OG in the blockchain space - would not be Chainmonsters if it did not incorporate those exciting aspects into Web3 as well.
As every Chainmon is trained individually, their trained model is going to be truly unique compared to any other Chainmon out there. This further enhances the opportunities provided through e.g. trading with other players as minted NFTs can be sold and bought on the Marketplace.
Some players who are really mastering both the game mechanics as well as seeking to optimize their playstyle might be encouraged to get into a true trainer role where they utilize their unique skillset to perfect the training capabilities and take on improving and then selling a Chainmon on the market.
This process is further enhanced and optimized through other unique NFTs including but not limited to Land, Dojos and materials.
Both minted Chainmons as well as materials and resources like food, candies and key items exist as NFTs on blockchain. This creates a player-driven economy approach where the community has to work together to achieve goals faster together. Specialized roles are created naturally including trainers, farmers, crafters and creators of specialized challenges and playgrounds in a User-Generated environment including the Dojos and Arenas.
A limited amount of Arenas are spread across the game world and each can only be controlled by a single individual player at the same time. The controlling player defines the Squad and environmental conditions that should give them a slight edge in combat. Other players need to specialize their competing Squad to those unique conditions until they can take over and make up the new rules.
More information regarding Arenas can be found here.
We want to encourage and incentivize experimentation and offer a set of ever changing challenges to our players where their AI squad needs to defeat our very own trained Chainmons! The expectation is that these challenges will be harder initially and then after targeted training players can get an edge over them.
See Challenges for other examples where our ML models are being utilized.
